{"event":{"id":"evt-aws-p5-h100-20230726","dedupe_key":"h100_cloud_ga:2023-07-26:aws-makes-h100-p5-instances-generally-available","event_type":"h100_cloud_ga","title":"AWS makes H100 P5 instances generally available","summary":"AWS made EC2 P5 instances powered by NVIDIA H100 GPUs generally available.","occurred_at":"2023-07-26T00:00:00.000Z","published_at":"2023-07-26T00:00:00.000Z","observed_at":"2026-10-07T15:56:31.000Z","reconstructed_at":"2026-10-07T16:04:51.000Z","ingest_type":"live_world_scan","locations":null,"status":"verified","confidence":0.99,"metadata":{"actors":["AWS"],"what_happened_at_time":"AWS made EC2 P5 instances powered by NVIDIA H100 GPUs generally available.","evidence_at_time":"AWS's July 26 release states P5 instances with H100 GPUs were generally available in specified U.S. regions for generative AI and HPC.","affected_layers":["layer-compute","layer-chips","layer-networks","layer-software"],"change_kind":"h100_hyperscaler_availability","q2_continuity":"Q1 introduced managed AI supercompute services; Q2 did not materially change Compute.","retrospective_inference":"H100-class acceleration became a concrete hyperscaler offering.","uncertainty":"Regional availability and pricing constraints mean this was not cheap or unconstrained frontier compute."},"created_at":"2026-10-07T16:56:16.849Z","updated_at":"2026-10-07T16:56:16.849Z"},"artifacts":[{"id":"art-aws-p5-h100-20230726","dedupe_key":"ba71dfe7b24d6523f57070a9d7aee2f39de0e3715835058cde02bbcf27c72da3","canonical_url":"https://aws.amazon.com/about-aws/whats-new/2023/07/amazon-ec2-p5-instances-generative-ai-hpc-generally-available/","evidence_locator":"","artifact_type":"h100_cloud_ga","title":"AWS P5 H100 general availability","summary":"AWS made EC2 P5 instances with NVIDIA H100 GPUs generally available.","creator_entities":["AWS"],"released_at":"2023-07-26T00:00:00.000Z","content_hash":null,"metadata":{"historical_scope":"2023-Q3"},"created_at":"2026-10-07T16:56:16.827Z","updated_at":"2026-10-07T16:56:16.827Z","relation":"evidenced_by"}],"signals":[{"id":"sig-q3-h100-hyperscaler-availability","statement":"H100-class AI acceleration became generally available through multiple major hyperscale cloud providers while remaining expensive, regional and concentrated.","signal_type":"frontier_compute_cloud_availability","direction":"increasing","confidence":0.98,"epistemic_status":"supported_inference","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T16:04:51.000Z","created_at":"2026-10-07T16:56:16.861Z","updated_at":"2026-10-07T16:56:16.861Z"}]}